Here’s one major change coming down the road: long-term support (LTS) for Linux kernels is being reduced from six to two years.
[…]Why? Simple, Corbet explained: “There’s really no point to maintaining it for that long because people are not using them.” I agree. While I’m sure someone out there is still running 4.14 in a production Linux system, there can’t be many of them.
Another reason, and a far bigger problem than simply maintaining LTS, according to Corbet, is that Linux code maintainers are burning out. It’s not that developers are a problem. The last few Linux releases have involved an average of more than 2,000 programmers — including about 200 new developers coming on board — working on each release. However, the maintainers — the people who check the code to see if it fits and works properly — are another matter.
The longer LTS support windows were put in place mostly for embedded devices, and as Ars Technica explains, it’s Android in particular that is affected by this change.
Just another form a force obsolescence at work, Linux people should be outraged.
While Instagram, TicTok and Snapchat continue to tell people extreme wealth, waste and consumption are good, it’s becoming less and less socially acceptable to make perfectly good functional hardware obsolete simply because the software or firmware is a little tired. it almost feels like a weird stealthy opposition to the Right to Repair.
It’s unconscionable for my 5 or 6 year old hardware, full of rare earth materials, dug from the ground and purified at a very high cost of the environment, becomes obsolete after a few years because the Devs got bored maintaining the code they have been well paid to provide.
exactly right. I had to upgrade my phone recently, not because it was broken in any way, but because the manufacturer decided it would no longer provide updates to the software.
Sadly the reality of internet connected devices that hold so much personal and financial information, it becomes quite risky to continue to use.
Linux people should be outraged? Why?
“It’s unconscionable for my 5 or 6 year old hardware, full of rare earth materials, dug from the ground and purified at a very high cost of the environment, becomes obsolete after a few years because the Devs got bored maintaining the code they have been well paid to provide.”
Most of the devs doing the work are doing so on a volunteer basis. If they were paid, by you or a company you paid money to with a contract to continue support, then you can blame them. The waste is terrible but its not due to boredom, why aren’t YOU doing the maintenance on six year old linux code bases for free? You have just as much reason to be mad at yourself as those that are giving up on the work.
Bill Shooter of Bul,
You weren’t talking to me, but I’ve been there, done that.
As I see it, it’s a problem that we have so many devices that do the exact same thing yet require their own unique device drivers. I don’t see why we need so many drivers to accomplish, it’s all about petty IP disputes. I think back to the era of clones before custom drivers where hardware was designed to be compatible (graphics, sound cards, network cards, etc). But windows normalized the “everybody implements their own drivers” way of supporting hardware…and it really sucks.
In instances where standards exist, like usb webcams and storage, we have broad compatibility that works very well in practice! I don’t know if there’s a opening the industry can take to create these standards across the board, probably not. But it would result in a lot more compatibility and less maintenance work for everyone.
Just install a newer LTS-Kernel! It is just some old LTS version that nobody is using any more that won’t get updates.
What is your problem?
Again Google getting a free ride. ♂ they should have maintainers.
Sometimes I wonder if the workload on the kernel devs would be less if they were to move beyond the mailing list approach to communication, and this is coming from a guy with 25+ years of experience in the field who has been through all the fads. Tools like GitHub/GitLab have made our development experience and speed much better. The ease of getting to all relevant comments associated with a ticket/issue/RFC is much nicer in my opinion than reeling through emails. Seeing the diffs for merge requests is pretty darn convenient as well. I won’t even comment on the CI/CD functionality that is already there. I get they already have much of that infrastructure in place, but having everything in one place sure is convenient especially when you are trying to get new people involved. Just my $0.02 though.
Agreed, but they are a finicky lot prone to argument over the most basic of tools. So email its been for ever.
I would agree that email is the most basic tool, and that is the problem because when your tools only satisfy the lowest common denominator that means the individual has to put forth more effort to accomplish a task. I empathize with the kernel devs with all the work, but from an outsiders perspective, they don’t seem to want to change where the change very well could make their life better. Sorry, but if this old coot can change with the times then they should at least make an effort and try.
They already have patchworks and a bugzilla, it just needs to become the main interface instead of parsing the LKML.
Kernel maintenance is very difficult and grueling work, I don’t envy those that are doing it despite my deepest gratitude. I did maintain a kernel driver for a couple of years, forward porting it to newer kernels and that was the most stressful job I did. Do I really understand what I’m doing? And if I screw this up my company is going under…
Bill Shooter of Bul,
I did so as well, it can be a thankless job. Users are disappointed when it doesn’t work, but they don’t really appreciate you when it does work.
I do feel there’s a lot of room for optimization though. The refusal to adapt a stable ABI Linux clearly results in lots of work amplification for linux devs. At this stage of project maturity and with maintenance under strain, it really seems they should be focusing on ways to reduce unnecessary workload and optimize workflow. The community has been very stubborn to change though so who knows if this will ever happen.
I didn’t even have to deal with other users. We were basically the only customer using our patches, because the hardware vendor only made patches for enterprise kernels. No one else was dumb enough to use our set up. We learned the hard way why enterprise distros exist.
My question is: WHo is the biggest users of Linux by volume? Are they server users? Or Desktop? If it’s the ladder, I see their point as Desktop HW updates at a near daily basis, ie: the Gamer crowd and their needs, so the obsolescence would be in order. Server/Workstation HW are usually supported for about 5 years so LTS is in line with that and the older HW on the pure consumer Desktop level with a 90 day warranty IMO.
If they want less burden they can reduce the frequency of LTS kernels. Currently, if I am not wrong, they are released them each 12 months. They could release them each 18-24 months without the world falling apart.
Some interesting perspectives, the paid only give new, the old stuff is all volunteer, and the cost of obsolescence is irrelevant.
While I sympathise with “Bill Shooter of Bul” the store down the road selling me a new embedded gadget with it’s new kernel and new widgets is very real, all with thousands of hours and millions of lines of kernel development, either as new or more often as a customisation of the old. Yet maintaining the old as is seems unsatisfactory and unrewarding.
I bet those gadgets are way more populous than cars, and the dollars per kilo of product will be much much higher, and yet they are obsolete and unserviceable after just a few years. Imagine if Mercedes did the same?
It seems someone somewhere is shirking a responsibility, and I doubt it’s any of the volunteers, why are volunteers even needed!
is that hard for big companies to keep up with latest lts? jesus.
each lts has many many many improvements. why skip all those?
None of the coverage I’ve seen of this has asked who the intended user of all this unused backporting is. It’s been years since the LTS lifetime was increased to to 6 years, but none of the commercial long term supported kernels have started even using those as a base for their own changes. There are commercially supported kernels much older than the LTS branches.
What failed? Does the kernel’s “LTS” process not meet the needs of actual long-term users? If not, why didn’t any discussion take place a long time ago?
The Ars article says it was basically for Android, but reconciling that with “no users” would mean it’s all for show, just to have a theoretically updatable kernel.